Genome Characteristics of Facultatively Symbiotic Frankia Sp. Strains Reflect Host Range and Host Plant Biogeography

Downloaded from genome.cshlp.org on October 4, 2021 - Published by Cold Spring Harbor Laboratory Press Article Genome characteristics of facultatively symbiotic Frankia sp. strains reflect host range and host plant biogeography Philippe Normand,1 Pascal Lapierre,2 Louis S. Tisa,3 Johann Peter Gogarten,2 Nicole Alloisio,1 Emilie Bagnarol,1 Carla A. Bassi,2 Alison M. Berry,4 Derek M. Bickhart,2 Nathalie Choisne,5,6 Arnaud Couloux,6 Benoit Cournoyer,1 Stephane Cruveiller,7 Vincent Daubin,8 Nadia Demange,6 Maria Pilar Francino,9 Eugene Goltsman,9 Ying Huang,2 Olga R. Kopp,10 Laurent Labarre,7 Alla Lapidus,9 Celine Lavire,1 Joelle Marechal,1 Michele Martinez,9 Juliana E. Mastronunzio,2 Beth C. Mullin,10 James Niemann,3 Pierre Pujic,1 Tania Rawnsley,3 Zoe Rouy,7 Chantal Schenowitz,6 Anita Sellstedt,11 Fernando Tavares,12 Jeffrey P. Tomkins,13 David Vallenet,7 Claudio Valverde,14 Luis G. Wall,14 Ying Wang,10 Claudine Medigue,7 and David R. Benson2,15 1Université de Lyon, Unité Mixte de Recherche, Centre National de la Recherche Scientifique (UMR CNRS), 5557 Ecologie Microbienne, IFR41 Bio Environnement et Santé, Université Lyon I, Villeurbanne 69622 cedex, France; 2Department of Molecular and Cell Biology, University of Connecticut, Storrs, Connecticut 06279, USA; 3Department of Microbiology, University of New Hampshire, Durham, New Hampshire, 03824, USA; 4Department of Plant Sciences, University of California, Davis, California 95616, USA; 5l’Institut National de la Recherche Agronomique–Unité de Recherche en Génomique Végétale (INRA-URGV), 91057 Evry cedex, France; 6Genoscope, Centre National de Séquençage, 91057 Evry cedex, France; 7Genoscope,CNRS-UMR 8030, Atelier de Génomique Comparative, 91006 Evry cedex, France; 8Bioinformatics and Evolutionary Genomics Laboratory, UMR CNRS 5558, Université Lyon I, Villeurbanne 69622 cedex, France; 9DOE Joint Genome Institute, Walnut Creek, California 94598, USA; 10Department of Biochemistry & Cellular & Molecular Biology and The Genome Science & Technology Program, The University of Tennessee, Knoxville, Tennessee 37996, USA; 11Department of Plant Physiology, Umeå University, S-90187 Umeå, Sweden; 12Instituto de Biologia Molecular e Celular, Universidade do Porto, 4150-180 Porto, Portugal; 13Clemson University Genomics Institute, Clemson, South Carolina 29634, USA; 14Programa Interacciones Biológicas, Departamento de Ciencia y Tecnología, Universidad Nacional de Quilmes, Bernal B1876BXD, Argentina Soil bacteria that also form mutualistic symbioses in plants encounter two major levels of selection. One occurs during adaptation to and survival in soil, and the other occurs in concert with host plant speciation and adaptation. Actinobacteria from the genus Frankia are facultative symbionts that form N2-fixing root nodules on diverse and globally distributed angiosperms in the “actinorhizal” symbioses. Three closely related clades of Frankia sp. strains are recognized; members of each clade infect a subset of plants from among eight angiosperm families. We sequenced the genomes from three strains; their sizes varied from 5.43 Mbp for a narrow host range strain (Frankia sp. strain HFPCcI3) to 7.50 Mbp for a medium host range strain (Frankia alni strain ACN14a) to 9.04 Mbp for a broad host range strain (Frankia sp. strain EAN1pec.) This size divergence is the largest yet reported for such closely related soil bacteria (97.8%–98.9% identity of 16S rRNA genes). The extent of gene deletion, duplication, and acquisition is in concert with the biogeographic history of the symbioses and host plant speciation. Host plant isolation favored genome contraction, whereas host plant diversification favored genome expansion. The results support the idea that major genome expansions as well as reductions can occur in facultative symbiotic soil bacteria as they respond to new environments in the context of their symbioses. [The genome sequences for Frankia strains CcI3, ACN14a, and EAN1pec have been submitted to GenBank under accession nos. CP000249, CT573213, and AAII00000000, respectively.] Two very different groups of bacteria can form nitrogen-fixing 15 Corresponding author. root nodules on angiosperms: Gram-negative proteobacteria E-mail [email protected]; fax 860-486-4331. Article published online before print. Article and publication date are at http:// from several families, and high Mol% G+C Gram-positive acti- www.genome.org/cgi/doi/10.1101/gr.5798407. nobacteria in the family Frankiaceae. Nodulating proteobacteria 17:000–000 ©2007 by Cold Spring Harbor Laboratory Press; ISSN 1088-9051/07; www.genome.org Genome Research 1 www.genome.org Downloaded from genome.cshlp.org on October 4, 2021 - Published by Cold Spring Harbor Laboratory Press Normand et al. have symbiotic genes (nod genes) subject to horizontal transfer in obligate bacterial pathogens and symbionts (Mira et al. 2001; among ␣- and some ␤-Proteobacteria (Chen et al. 1991; Young Ochman and Moran 2001; Moran 2003), but the observation and Haukka 1996; Moulin et al. 2001). In contrast, all Frankia sp. that both contraction and expansion can occur in closely related strains are closely related with no evidence of dissemination of lineages of facultatively symbiotic soil bacteria in relation to host nodulating ability to related actinobacteria (Fig. 1; Normand et distribution has not previously been reported. al. 1996; Clawson et al. 2004). In plants, the capacity to form N2-fixing root nodules occu- pied by bacteria is retained in a single lineage of angiosperms Results and Discussion known as the “N -fixing clade” (Soltis et al. 1995). Ten families 2 ∼ within the Eurosid I clade have members that are nodulated Actinorhizal plant families emerged in the late Cretaceous ( 100 (Soltis et al. 1995; Swensen 1996; Clawson et al. 2004). Only two million years ago [Mya]) and subsequently adapted to a wide of the families have members that associate with nodulating pro- variety of environments (Magallon et al. 1999). Currently, they teobacteria, while eight associate with Frankia sp. strains to form are globally distributed in climate zones ranging from alpine and the actinorhizal symbiosis (Table 1). subarctic to tropical (Fig. 2) where they add nitrogen and organic Frankia strains fall into three closely related clusters. Mem- material to nutrient-poor soils (Silvester 1976). The native geo- bers of each cluster have distinct host ranges (Table 1; Fig. 1). graphical distributions of hosts range from limited in the case of Cluster 1 strains nodulate plants in the Fagales in the Betulaceae Casuarina sp. to broad in the case of Morella sp. (Fig. 2). The and Myricaceae and are often refered to as “Alnus strains” (Nor- distribution of bacterial symbionts is obviously more difficult to mand et al. 1996). A subclade within Cluster 1 is comprised of the assess, but numerous studies have shown some correlation with narrow host range “Casuarina strains” that under natural condi- plant distribution (for review, see Benson et al. 2004). tions nodulate only Casuarina and Allocasuarina species in the Frankia sp. strain HFPCcI3 (CcI3) represents narrow host Casuarinaceae (Benson et al. 2004). Conversely, Cluster 3 “Elae- range Casuarina strains commonly detected in nodules collected agnus strains” are considered to have a broad host range since from casuarinas in their native Australia (Fig. 2A) and in areas of they nodulate plants from five families in the Fagales and Rosales the world where casuarina trees have been planted as windbreaks (Benson et al. 2004). Finally, the “Rosaceous strains” form Clus- or for erosion control (Simonet et al. 1999). Similar strains have ter 2, which is sister to the others; representatives of this cluster not been found in soils in the absence of a suitable host, indi- have not been isolated and grown in culture. Cluster 2 strains cating that the bacteria depend on the plant for their soil propa- nodulate plants from four families in the Rosales and Cucurbi- gation (Simonet et al. 1999). tales (Benson et al. 2004; Vanden Heuvel et al. 2004). Frankia alni strain ACN14a (ACN) represents Alnus strains To gain insight into the evolutionary trajectory followed by that are globally distributed in soils regardless of the presence of these closely related, yet host-range and geographically diver- a suitable host plant (Benson et al. 2004). This ubiquity parallels gent, Frankia sp. strains, we sequenced and compared the ge- the distribution of host plants from the Betulaceae and Myrica- nomes of three isolates, including a narrow host range Casuarina ceae that have a combined native range spanning all continents strain, a medium host range Alnus strain, and a broad host range except Australia (Table 1; Fig. 2B). Elaeagnus strain. The results suggest that gene deletion and du- Frankia sp. strain EAN1pec (EAN) represents broad host plication have occurred to different extents in the genomes dur- range Elaeagnus strains that are also globally distributed in soils ing adaptation to host plants and their environments. The con- with or without host plants (Benson et al. 2004). Cognate hosts cept of genome contraction echoes the changes known to occur are the most diverse and have the widest distribution with representatives on all continents including Australia (Table 1; Fig. 2C). The strains used in this study have 16S rRNA gene sequences that are 97.8% identical between ACN or CcI3 versus EAN, and 98.9% identical between ACN and CcI3 (Fig. 1). This similarity level is frequently observed among bacteria from the same species (Wayne et al. 1987; Gevers et al. 2005), and is typical of the similarity levels found within the genus Frankia (Fig. 1; Clawson et al. 2004). Genome characteristics The genomes from ACN and CcI3 have been finished, and that from EAN has been rendered in a single scaffold with some gaps corresponding to regions that have proven difficult to resolve due to sequence repeats and high GC content (Table 2). Never- theless, unlike Streptomyces (Bentley et al. 2002), all three ge- Figure 1. Neighbor-joining (Saitou and Nei 1987) phylogenetic tree nomes are circular as demonstrated directly from their sequences calculated with ClustalX 1.83 (Thompson et al.

Load more